Topic: Hanger! Table buffer for alias... contains uncommitted changes.

Help, and that a hanger!
I anywhere did not adjust buffering since this I do not use, but for quite some time now on some cursors this began to torment me ! And in identical conditions one cursors accurately work without buffering (for a long time the created forms), and nothing remarkable others (recently created forms) - give an error.
All cursors are samplings with SQL-serva through SQLEXEC () and are sources for  - and on the typical form (a class, everywhere same, without imported changes). On a course this is real all reference manuals. I there in a class of the form such made a button - to staticize the data on which in  the data is overloaded anew, in the same cursor, without its (cursor) of preliminary closing. After in such cursor happens any  and after that is pushed to "Staticize" and there is "miracle" - one cursors reboot the data normally, and others - fall the program. If  had no place no error arises.
What to do, and why such  and   to happen?
. CURSORGETPROP (' Buffering ') for some reason always gives 3 though on-default  to give 1. I  shock, but not from that, and that that at attempt CURSORSETPROP (' Buffering ', 1) me produce - Views require either DB_BUFOPTROW or DB_BUFOPTTABLE. At what  Views? I them in eyes never , not that though time though somewhere used that! A fuck-up full...


Re: Hanger! Table buffer for alias... contains uncommitted changes.

I will not understand one - why about one year in similar situations the problem did not float. How to overcome consequences - me it is clear. The matter is that to close the cursor it is possible without problems and for some reason  not against. I would like to learn the reason. Understand, all is identical, only really commands sent on the server is separated, well and fields on names, by itself. It is almost unique difference and nevertheless, one in absolutely identical conditions work without errors, and others -  an error in line with SQLEXEC (nH, cCmd, cCursor) because does not wish  in the same, open cursor, given to flood.
You understand, if the error takes place, it is always if conditions for its origin form. And here it turns out .
I why, actually, the inhuman text - am necessary tomorrow indicative demonstration, on a course even for today was planned, and here in a place where all perfectly worked suddenly, understand, such...
So, all. I seem understood in what business...
Thanks for education in sense:


the Cursor received through SQLExec () always is in a mode of optimistic buffering. By default, in a mode of optimistic buffering of lines (3), but it is possible to switch in a mode of optimistic buffering of tables (5).
To switch such cursor in a mode of pessimistic buffering or to disconnect for it an optimization mode - it is impossible.

did not know.
On-mojmu the reason that by means of SQLExec () I do not use the majority of the received cursors, and I use the cursor received from the subsequent sampling of it. Therefore, seemingly, it till now also did not face never... If not you, I do not know when I would reach this thought! Thanks!